Project Summary Small cell lung cancer (SCLC) afflicts more than 30,000 patients per year and is rapidly fatal in 95% of cases, with median survival is less than one year. Belying this grim prognosis, treatment-naive SCLC is highly sensitive to chemotherapy. However, relapse is nearly inevitable, and relapsed SCLC presents two obstacles that have been insurmountable for at least 30 years: cross-resistance to chemotherapy, and absence of biomarker-driven targeted therapy. Following relapse, resistance often extends beyond etoposide/platinum (EP) to other DNA damaging agents. Although topotecan is the only approved second-line therapy for SCLC, the NCCN guidelines list 10 agents of roughly equivalent efficacy. None are particularly effective in unselected patients, and a disease that was once highly chemosensitive becomes inexorably progressive. However, the molecular determinants of cross- resistance in SCLC remain unclear. Although critically important, cross-resistance is difficult to study experimentally, as it requires a model system that faithfully reproduces clinical outcomes, and is adequately powered to capture inter-tumoral molecular heterogeneity across a population of patients. We have generated a panel of 44 SCLC patient-derived xenograft models (PDXs) from biopsy specimens and circulating tumor cells (CTCs). Our panel includes successive models from individual patients at time points before and after specific lines of therapy, with detailed information about the corresponding clinical response. For both standard chemotherapy and experimental agents in clinical trial, these models faithfully mirror patient responses. However, unlike the patient experience, multiple strategies can be compared for identical tumors. We propose to use these models to directly compare three clinical strategies that depend on induction of DNA damage: standard first line EP, second line topotecan, anad a promising experimental regimen, olaparib plus temozolomide (OT), currently in a phase I/II trial at MGH. Individually, these PDX population trials are designed to reveal biomarkers of sensitivity and mechanisms of resistance for promising experimental therapies. Collectively, through comparative analysis with reference to the clinical histories of each model, they present a novel opportunity to model cross-resistance, a problem that has beleaguered management of SCLC for over three decades.